Mechjeb: Realism or not?

[juanml82]

I suppose is likely that you can't pilot any rocket or space craft manually IRL, there's always an autopilot or guidance system, however, NASA doesn't fly rockets in an universe that can be described with just a bunch of simple deterministic formulas.

Well, real astronauts also have a far steeper learning curve than some random guy playing a videogame

*Sigh*

Can we read the precise location of where we are in the universe now, or do we calculate it with simplistic approximations and the correct via manual or instrument input?

Do we have the on-board AI execute a series of maneuvers by analyzing "the universal source code", or are they limited by the instruments on board? (Akin to how ISA-MapSat knows the geo-position without a single radar station to say "yep, you're exactly there")

I'd still doubt that NASA wouldn't perform course trajectory on their own N-Body Supercomputer than use approximations for an Autopilot. Even then, power constraints asks what is so difficult with programming in a flight plan that you need the extra power for it to be calculated and adjusted in real-time.

Be realistic here, "Autopilot" has far too many meanings to just say "NASA Uses Autopilots, QED MechJeb is Realistic". And MechJeb's ability to get utterly precise information, in real time, and make calculations on that, in real time, is a little absurd. Yes, we don't collect mun rocks, or have waste, or do anything to change the mass of the ship... but there are differences between a precise figure read from "the magic of the universe" and a calculated figure that has some error to it... and how small error builds up and eventually requires outside observation to correct for.

Whatever really. Whatever.

*In short*

Real autopilots use assisting technology to perform precision maneuvers. Without GPS and Radar there would BE no "Aircraft Autopilot", and similar approaches should be taken when talking of a space autopilot.

AFAIK, space ships (and ICBMs) use the stars to navigate - the same as the 19th century sailors, but with atomic clocks and computers. And that way, it is precise.

[Red Iron Crown]

AFAIK, space ships (and ICBMs) use the stars to navigate - the same as the 19th century sailors, but with atomic clocks and computers. And that way, it is precise.

GPS or equivalent these days, so we've substituted an artificial constellation for the natural ones.

[juanml82]

GPS or equivalent these days, so we've substituted an artificial constellation for the natural ones.

GPS doesn't work for probes beyond Earth orbit. As for ICBMs, I think they don't use GPS because they can be jammed and, in the event of a nuclear war, the GPS satellites are targets. The stars are far more reliable and, above the stratosphere, they are always visible.

[Red Iron Crown]

GPS doesn't work for probes beyond Earth orbit. As for ICBMs, I think they don't use GPS because they can be jammed and, in the event of a nuclear war, the GPS satellites are targets. The stars are far more reliable and, above the stratosphere, they are always visible.

GPS works out as far as GEO and perhaps further, though of course anything deeper in space than that will not use it. Most of the US' missiles and other PGMs have migrated to GPS, but it's possible I'm wrong about ICBMs using it (likely so now that I think about it, there hasn't been a new design by the US in quite some time).

[undercoveryankee]

More specifically, a star tracker tells you your orientation relative to the background stars, which are far enough away that their directions are basically parallel anywhere in the solar system. To get position, you need angles or distances to objects within the solar system. A sailor with a sextant uses angles between the Sun, Moon, and stars and Earth's horizon, along with the most accurate clock he can afford to carry, to find the location on Earth where gravity points the right direction relative to the stars he can see.

If you were navigating a spacecraft without help from the ground, you or your instruments would be doing basically the same thing, using measurements like "angle between a star and the nearest edge of a nearby planet or moon" or "time when a star crosses the limb of a planet." The time when you gain or lose signal between the spacecraft and a particular relay station adds a valuable data point; for spacecraft doing burns on the far side of the moon, the time of reacquisition of signal was often the first indication the ground had of whether the burn had gone correctly. The ground has radars, telescopes, and better clocks, and whenever you have line of sight probably knows more about your position than you do.

Since manual star sightings are slow and even automated star trackers can take a while to get their bearings after a major maneuver, spacecraft use various types of gyroscopes to maintain their orientation between sightings and correct any errors in real time. A modern inertial measurement unit also includes sensitive accelerometers to keep track of the actual acceleration the vehicle is experiencing. This is essential if you're in atmosphere and have to correct for wind, turbulence, or variations in engine performance moving you around. In space, many engines are predictable enough that burning for a calculated time is enough to achieve the necessary delta-v accuracy. If you're designing a vehicle where that's the case, you can use lighter or less expensive accelerometers because their purpose becomes catching equipment malfunctions.

There are some cases in games where adding more realism to what's being simulated can make the outcomes less realistic. Yes, there's software available for real spacecraft to do everything that MechJeb offers, if not on the vehicle itself than by uplinking the results of calculations done on the ground. But there are tough design decisions about what sensors and how much drive space for software to put on a vehicle, so real spacecraft typically don't have the flexibility to command any maneuver you can design at any time.

MechJeb by itself gives you the software that makes your life easier than stock KSP, while abstracting over the design decisions that are needed to make those capabilities work, so it's reasonable to find that hand-flying is a better-balanced game. It's one of those cases where making one aspect more realistic, without simulating other things that balance it, can give you a less realistic overall outcome. I use it so I can spend more time being creative and doing the interesting parts of missions and less time punching numbers into spreadsheets and flying routine mission profiles to lay the groundwork for the next interesting thing.

[undercoveryankee]

GPS works out as far as GEO and perhaps further, though of course anything deeper in space than that will not use it. Most of the US' missiles and other PGMs have migrated to GPS, but it's possible I'm wrong about ICBMs using it (likely so now that I think about it, there hasn't been a new design by the US in quite some time).

Missiles probably have it, as part of a philosophy of using whatever sources of information are working at any given moment. Civilian GPS receivers are required to cut out if they exceed specified speed and altitude limits so they can't be used to guide missiles off the shelf.

[NathanKell]

Most ICBMs had inertial guidance back in the day; I don't know what they have now.

[Capt. Hunt]

even modern JDAM bombs use INS, it's simple, accurate and cheap It just uses the GPS to get the initial coordinate data, I'm pretty sure modern ICBMs and cruise missiles use a similar system. Also, remember that the early mercury flights were all remote guided, Alen Shepard didn't have any more control of the rocket then the Chimp they sent up before him. I see nothing wrong with using Mechjeb, as long as you figure out how to do it yourself first, which is why mechjeb's features are tied to the tech tree anyway.

[Axelord FTW]

It's realistic-ish.

That said, I would also tell anyone getting into the game to steer clear of MJ early on.

Really, players should try and do everything manually at least once. I had let go of KSP for a good year before getting back on it and I made a point to make my first career play stock and I can definitely say my experience was better. As a gameplay element, MJ is casual but it actually helps trudging through repeated gameplay.